[rrd-users] 5 second resolution definition.

Simon Hobson linux at thehobsons.co.uk
Fri Sep 19 15:28:21 CEST 2008

<Gerhardus.Geldenhuis at gta-travel.com> wrote:

>I am trying to define an RRA for my graph but I do not understand how to
>define the "part" section of the RRA definition

"part" section ?

>5 minutes = 0.5
>1 minute = 0.25
>However I dont understand how this value is determined. I do not
>understand how 5 minutes can be 0.5 and 1 minute be 0.25. I can't see
>any relation so I dont know how to deduce anything further from it.
>Googling has only turned up 5 and 1 minute examples so far.


I think you are getting confused.

step sizes are defined in seconds - so "5" for 5 seconds in the rrd definition.

Then you need to define one or more DSs, and here you need to define 
"heartbeat" - how long you can go between samples before the data is 
deemed to be invalid. Again, this is defined in seconds.

And lastly, define one or more consolidation functions (CF). I thik 
this is where you are getting confused, as "0.5" and "0.25" sound 
like xff values : "xff The xfiles factor defines what part of a 
consolidation interval may be made up from *UNKNOWN* data while the 
consolidated value is still regarded as known." A value of 0.5 means 
that HALF of the values in the consolidation interval may be unknown 
before the whole interval is considered unknown.

More information about the rrd-users mailing list